Cross-domain EEG-based Emotion Recognition with Contrastive Learning
This work addresses robust emotion recognition for affective computing applications, showing strong performance but being incremental as it adapts an existing framework to a specific domain.
The paper tackled cross-domain EEG-based emotion recognition by reformulating it as an EEG-text matching task using the CLIP framework, achieving superior cross-subject accuracies of 88.69% and 73.50% and cross-time accuracies of 88.46% and 77.54% on SEED and SEED-IV datasets.
Electroencephalogram (EEG)-based emotion recognition is vital for affective computing but faces challenges in feature utilization and cross-domain generalization. This work introduces EmotionCLIP, which reformulates recognition as an EEG-text matching task within the CLIP framework. A tailored backbone, SST-LegoViT, captures spatial, spectral, and temporal features using multi-scale convolution and Transformer modules. Experiments on SEED and SEED-IV datasets show superior cross-subject accuracies of 88.69% and 73.50%, and cross-time accuracies of 88.46% and 77.54%, outperforming existing models. Results demonstrate the effectiveness of multimodal contrastive learning for robust EEG emotion recognition.